摘要
为了提高位置社交网络的服务便捷性和用户感受度,与位置相关的推荐服务越来越具有重要意义和应用需求。提出的潜在好友推荐模型主要是根据签到位置的相似度及好友相似度进行潜在用户推荐。通过用户的好友关系、签到特性及签到历史记录,计算用户在各个位置兴趣点的位置权重,再分别利用位置权重及好友关系计算用户的位置相似度和好友相似度,最后根据用户位置和好友关系的综合相似度进行潜在用户推荐。实验结果表明,提出的潜在好友推荐模型是切实有效的。
To make service convenient and improve user experience degrees, recommendation service has becomemore and more important to users in the location-based social network （LBSN）. An improved potential friendsrecommendation in LBSN, which is based on the relationship among friends and the similarity of locations, wasproposed. First of all, the user interest pointsin each position weight with friends＇ relationship, check-in feature andhistorical recordwere calculated. Then position weight and the friend relationship were used to calculate the friends＇similarity and location similarity. Finally, potential friends to user based on the friends＇ similarity and locationsimilarity were recommended. The experimental results show that the proposed model is effective.
出处
《电信科学》
北大核心
2014年第10期71-77,共7页
Telecommunications Science
基金
国家自然科学基金资助项目(No.61370139)
网络文化与数字传播北京市重点实验室基金资助项目(No.ICDD201309)
北京市属高等学校创新团队建设与教师职业发展计划基金资助项目(No.IDHT20130519)
关键词
位置服务
位置社交网络
潜在好友
推荐模型
位置相似度
location-based service, location-based social network, potential friend, recommendation model, locationsimilarity
